This new R01 proposal is a collaboration between two investigators, Wittrup and Irvine, combining protein engineering and nanoparticle synthesis expertise. Our central hypothesis is that the therapeutic index of cancer immunotherapy can be improved significantly by using novel methods to locally concentrate potent immunostimulatory molecules in tumor tissue for increased efficacy and decreased off-target toxicity. We will develop two complementary and potentially synergistic localized delivery methods for immunotherapy of cancer: pretargeting, and intratumoral nanoparticle injection. The two methods will be optimized for combined utility in syngeneic and genetically engineered mouse tumor models. We will explicate the immune therapeutic mechanisms of protocols that demonstrate efficacy. We have developed a bispecific antibody-based pretargeting protocol that provides highly tumor-specific localization of the chelator DOTA. We will site-specifically attach DOTA to the payloads IL-2, IL-12, TNF-, -CTLA4 scFv, and -CD137 scFv. These molecules were chosen due to their demonstrated immuno-therapeutic potential in clinical trials, together with significant toxicity issues. Our protocol validated for DOTA-radiometal chelates will be adapted for specific delivery of the DOTA-labeled payloads. We have devised liposomal and stabilized micellar vehicles for surface anchoring of immunostimulatory molecules, and demonstrated their efficacy and safety from intratumoral injection into B16F10 syngeneic melanoma tumors. The same bispecific antibody used for pretargeting will be anchored on the surface of these vehicles, so that the exact same DOTA-labeled payloads can be modularly tested without re-optimization of conjugation methods. The bsAb is a scaffold that enables straightforward mimicry of immunocytokines, bispecific antibodies, and Fc conjugates by noncovalent conjugation with DOTA-labeled payloads. This will enable us to benchmark safety and efficacy of our novel approaches against these more commonly used vehicles, using the same antibody for tumor targeting and identical immunostimulatory molecules. We will test these protocols in transgenic mice expressing CEA, inoculated subcutaneously with B16F10 tumors expressing human CEA. The most successful protocols will be further tested in subcutaneous MC38-CEA tumors, and then in genetically engineered KP tumors in lung and sarcoma (floxed p53 knockout and stop-floxed activated KRAS expression via Cre recombinase delivered virally.) We will closely examine the tumor microenvironment and tumor draining lymph nodes following treatment by the most efficacious protocols, for evidence of reversal of immunosuppression by Tregs, TAMs, or MDSCs. We will also test for protective immunity and antigen spreading using syngeneic tumors lacking the antigen targeted by the bsAb (CEA).